A Habitable Fluvio-Lacustrine Environment at Yellowknife Bay, Gale Crater, Mars J. P. Grotzinger et al. Science 343, (2014); DOI: 10.1126/science.1242777

This copy is for your personal, non-commercial use only.

If you wish to distribute this article to others, you can order high-quality copies for your colleagues, clients, or customers by clicking here. Permission to republish or repurpose articles or portions of articles can be obtained by following the guidelines here.

The following resources related to this article are available online at www.sciencemag.org (this information is current as of February 10, 2014 ):

Updated information and services, including high-resolution figures, can be found in the online

version of this article at: on February 10, 2014 http://www.sciencemag.org/content/343/6169/1242777.full.html Supporting Online Material can be found at: http://www.sciencemag.org/content/suppl/2013/12/05/science.1242777.DC1.html This article cites 72 articles, 32 of which can be accessed free: http://www.sciencemag.org/content/343/6169/1242777.full.html#ref-list-1 This article has been cited by 7 articles hosted by HighWire Press; see:

http://www.sciencemag.org/content/343/6169/1242777.full.html#related-urls www.sciencemag.org This article appears in the following subject collections: Planetary Science http://www.sciencemag.org/cgi/collection/planet_sci Downloaded from

Science (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by the American Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. Copyright 2014 by the American Association for the Advancement of Science; all rights reserved. The title Science is a registered trademark of AAAS. RESEARCH ARTICLE

oped (7), along with Gale’s relevance to MSL’s goals (1, 8). The primary mission objective is to examine the lower foothills of Mt. Sharp (Fig. A Habitable Fluvio-Lacustrine 1A), located outside of the landing ellipse, where hematitic and hydrated clay- and sulfate-bearing Environment at Yellowknife Bay, strata were detected from orbit (5, 9) and are accessible by the rover (1). The surface mission was planned in ad- Gale Crater, Mars vance (1) to take into account the drive dis- tance from anywhere in the landing ellipse up 1 2 3 1 4 5 6 J. P. Grotzinger, * D. Y. Sumner, L. C. Kah, K. Stack, S. Gupta, L. Edgar, D. Rubin, † to and across the Mt. Sharp foothills, includ- 7 8 9 10 11 12 5 K. Lewis, J. Schieber, N. Mangold, R. Milliken, P. G. Conrad, D. DesMarais, J. Farmer, ing time for sampling and analyses en route. 1 13 14 14 15 16 13 K. Siebach, F. Calef III, J. Hurowitz, S. M. McLennan, D. Ming, D. Vaniman, J. Crisp, The landing ellipse is located in the broad val- 13 17 17 12 18 11 19 A. Vasavada, K. S. Edgett, M. Malin, D. Blake, R. Gellert, P. Mahaffy, R. C. Wiens, ley (Aeolis Palus) between the Gale crater rim 20 21 21 13 13 22 23 S. Maurice, J. A. Grant, S. Wilson, R. C. Anderson, L. Beegle, R. Arvidson, B. Hallet, and Mt. Sharp (Fig. 1A). Depending on where R. S. Sletten,23 M. Rice,1 J. Bell III,5 J. Griffes,1 B. Ehlmann,1,13 R. B. Anderson,24 T. F. Bristow,12 W. E. Dietrich,25 G. Dromart,26 J. Eigenbrode,11 A. Fraeman,22 C. Hardgrove,17 K. Herkenhoff,24 1 L. Jandura,13 G. Kocurek,27 S. Lee,13 L. A. Leshin,28 R. Leveille,29 D. Limonadi,13 J. Maki,13 Division of Geologic and Planetary Sciences, California In- stitute of Technology, Pasadena, CA 91125, USA. 2Depart- S. McCloskey,13 M. Meyer,30 M. Minitti,5 H. Newsom,31 D. Oehler,15 A. Okon,13 M. Palucis,25 13 32 33 34 35 1 36 ment of Earth and Planetary Sciences, University of California, T. Parker, S. Rowland, M. Schmidt, S. Squyres, A. Steele, E. Stolper, R. Summons, Davis, CA 95616, USA. 3Department of Earth and Planetary A. Treiman,37 R. Williams,16 A. Yingst,16 MSL Science Team‡ Sciences, University of Tennessee, Knoxville, TN 37996, USA. 4Department of Earth Science and Engineering, Imperial College London, London SW7 2AZ, UK. 5School of Earth and The Curiosity rover discovered fine-grained sedimentary rocks, which are inferred to represent an Space Exploration, Arizona State University, Tempe, AZ 85287, ancient lake and preserve evidence of an environment that would have been suited to support a USA. 6U.S. Geological Survey, Santa Cruz, CA 95060, USA. martian biosphere founded on chemolithoautotrophy. This aqueous environment was characterized 7Department of Geoscience, Princeton University, Princeton, 8 by neutral pH, low salinity, and variable redox states of both iron and sulfur species. Carbon, NJ 08544, USA. Department of Geological Sciences, Indiana University, Bloomington, IN 47405, USA. 9Laboratoire Planétologie hydrogen, oxygen, sulfur, nitrogen, and phosphorus were measured directly as key biogenic et Géodynamique de Nantes (LPGN), LPGN/CNRS UMR6112 elements; by inference, phosphorus is assumed to have been available. The environment probably and Université de Nantes, 44322 Nantes, France. 10Depart- had a minimum duration of hundreds to tens of thousands of years. These results highlight the ment of Geological Sciences, Brown University, Providence, RI 11 biological viability of fluvial-lacustrine environments in the post-Noachian history of Mars. 02912, USA. NASA Goddard Space Flight Center, Greenbelt, MD 20771, USA. 12Department of Space Sciences, NASA Ames Research Center, Moffett Field, CA 94035, USA. 13Jet Propul- he environmental history of early Mars included well-defined secondary objectives in sion Laboratory, California Institute of Technology, Pasadena, is mostly written in stone. The first few addition to exploring Gale’s central mountain, CA 91109, USA. 14Department of Geosciences, State University billion years of Mars’ geologic history Aeolis Mons (informally known as Mt. Sharp), of New York State at Stony Brook, Stony Brook, NY 11794– T 2100, USA. 15Jacobs Technology, NASA Johnson Space Center, ’ document surface environments considerably the mission s primary science target. The selec- Houston, TX 77058, USA. 16Planetary Science Institute, Tucson, different than the surface today, prompting a tion of the Gale crater field site, where Curiosity AZ 85719, USA. 17Malin Space Science Systems, San Diego, CA succession of coordinated surface and orbiter landed on 6 August 2012, 5:32 UTC, was based 92121, USA. 18Department of Physics, University of Guelph, missions over the past two decades aimed at on the recognition that it retained records of Guelph, Ontario N1G 2W1, Canada. 19Space Remote Sensing, ultimately determining if Mars ever had an early multiple and diverse ancient aqueous environ- Los Alamos National Laboratory, Los Alamos, NM 87544, USA. 20Institut de Recherche en Astrophysique et Planétologie (IRAP), biosphere. Water is the essential ingredient for ments and thus enhanced the potential that one Université de Toulouse/CNRS, 31400 Toulouse, France. 21Center all life as we understand it; therefore, past mis- or more of those settings might have provided for Earth and Planetary Studies, National Air and Space Museum, sions have sought environments in which water the combination of factors necessary to sustain a Smithsonian Institution, Washington, DC 20560, USA. 22Depart- was abundant and possibly long-lived. How- habitable environment (1). Another important fac- ment of Earth and Planetary Sciences, Washington University in St. Louis, St. Louis, MO 63130, USA. 23Department of Earth ever, to create an environment that could have tor included mapping the landing ellipse ahead of and Space Sciences, University of Washington, Seattle, WA been habitable by microorganisms, the pres- landing so that no matter where the rover touched 98195, USA. 24U.S. Geological Survey, Flagstaff, AZ 86001, ence of water alone is insufficient: A source of down, our first drive would take us in the direc- USA. 25Department of Earth and Planetary Science, University 26 energy to fuel microbial metabolism is required, tion of a science target deemed to have the greatest of California, Berkeley, CA 94720, USA. Laboratoire de Geologié de Lyon, Université de Lyon, 69364 Lyon, France. 27Department as are the elements carbon, hydrogen, sulfur, ni- value, as weighed against longer-term objectives of Geological Sciences, University of Texas at Austin, Austin, trogen, and phosphorous (and a host of others at and the risk of mobility failure. TX 78712, USA. 28School of Science, Rensselaer Polytechnic trace levels). The Mars Science Laboratory (MSL) Institute, Troy, NY 12180, USA. 29Canadian Space Agency, 30 rover Curiosity was designed and built to search Mapping and Geologic Framework St-Hubert, Quebec J3Y 8Y9, Canada. NASA Headquarters, for Exploration Washington, DC 20546, USA. 31Institute of Meteoritics, Univer- for these materials and, thus, potentially delin- sity of New Mexico, Albuquerque, NM, 87131 USA. 32Depart- eate one or more habitable environments at the Gale crater is 154 km in diameter with a cen- ment of Geology and Geophysics, University of Hawaii at Manoa, Gale crater landing site. We show here that Gale tral mountain of stratified rock. Gale straddles Honolulu, HI 96822, USA. 33Department of Earth Sciences, Brock crater once contained an ancient lake that would the crustal dichotomy boundary, a topographic University, St. Catharines, Ontario L2S 3A1, Canada. 34Depart- have been well suited to support a martian bio- demarcation between heavily cratered uplands ment of Astronomy, Cornell University, Ithaca, NY 14853, USA. 35Geophysical Laboratory, Carnegie Institution for Science, sphere founded on chemolithoautotrophy. and sparsely cratered lowlands. In the vicin- Washington, DC 20015, USA. 36Department of Earth and Plan- Exploration for habitable environments is ity of Gale, the dichotomy boundary is crossed etary Science, Massachusetts Institute of Technology, Cambridge, Curiosity’s core mission objective. Achieving by numerous incised valley networks sugges- MA 02139, USA. 37Lunar and Planetary Institute, Houston, this goal necessitates a challenging set of sci- tive of surface aqueous flows that discharged TX 77058, USA. ence measurements on just the right rocks to to the north across the lowlands (2). Gale’sstra- *Corresponding author. E-mail: [email protected] †Present address: Department of Earth and Planetary Sciences, extend beyond the search for aqueous environ- tigraphy, mineralogy, and landforms have been University of California, Santa Cruz, CA 95064, USA. ments. To enhance the probability of mission well studied from orbit (3–6). A broader context ‡MSL Science Team authors and affiliations are listed in success, we adopted an exploration strategy that for understanding Mt. Sharp has been devel- the supplementary materials.

www.sciencemag.org SCIENCE VOL 343 24 JANUARY 2014 1242777-1 Exploring Martian Habitability Curiosity touched down, potential secondary sci- thermal inertia exposed at Yellowknife Bay (11) presence of scarps and lineations consistent with ence targets within the landing ellipse included (Figs. 1A and 2A). bedding; surface roughness based on shadow- the Peace Vallis alluvial (water-deposited) fan Geologic mapping (12), building on earlier ing; apparent relative albedo; and patterns of lobe; a high–thermal inertia terrain localized efforts (4), proved critically important to se- albedo variation such as those indicating frac- downslope from and adjacent to the fan; sev- lecting the postlanding drive direction and sci- tures, bedding, or mottling. These units include eral fresh craters; and inverted stream channels. ence targets. This work provided the broader the Peace Vallis alluvial fan (AF); an immediate- In the case of the current study, a secondary target context in which to understand the strata drilled ly adjacent and downslope light-toned, bedded, near the landing site became a notable objective, by Curiosity in Yellowknife Bay. The MSL team– fractured unit (BF); surfaces with relatively high with the decision (10) made after landing to drive generated map (Fig. 1B) reveals six units (13) crater densities (CS); tonally smooth but hum- in the opposite direction of the Mt. Sharp entry that are distinguished based on variations in mocky plains (HP); light-toned topographically point to sample stratified rocks of relatively high thermal inertia; textural attributes such as the variable or rugged terrain (RT); and light-toned

Fig. 1. Regional context maps. (A)LocationofBradbury landing site and Yellowknife Bayinrelationtothetopog- raphy and thermal inertia of the broad valley between Gale crater’s rim and Mt. Sharp. The white ellipse denotes the landing ellipse; the white line represents the intended drive route from Yellowknife Bay to the Mt. Sharp entry pointatMurrayButtes.(B)A geological map was constructed based largely on HiRISE im- age data and was used to demarcate major terrain types for exploration by Curiosity. The MSL Science Team chose to drive the rover to Yellow- knife Bay where three terrains intersect in a triple junction. One of these terrains (BF) very closely coincides with higher values of thermal inertia, and it was also recognized that rocks at Yellowknife Bay are downslope of those forming the AF unit, suggesting a ge- netic link. The black ellipse is the landing ellipse; the white line represents the intended drive route from Yellowknife Bay to the Mt. Sharp entry point. The legend for sym- bols in (A) also applies to those in (B).

1242777-2 24 JANUARY 2014 VOL 343 SCIENCE www.sciencemag.org RESEARCH ARTICLE striated rocks (SR). By sol 440 (where 1 sol is Landing, including the excavation zones produced aligned linear ridges with a relief of 0.5 to 2.5 m a martian day), the RT and SR units had not by Curiosity’s descent rockets, expose pebble interpreted as inverted stream channels. Down- been examined by Curiosity and are not dis- conglomerate bedrock of fluvial (stream-related) slope, this unit passes into the widespread BF cussed further. Both are considered as expo- origin (17). unit that continues all the way to Yellowknife sures of bedrock based on textures and higher The AF and BF units are important because Bay, where it is coincident with the stratified thermal inertia values. The CS unit was inspected of their intimate spatial association and possi- bedrock exposed there and the terrain that marks only from a distance during the traverses to and ble relation to rocks exposed at Yellowknife the floor of the “bay” (18) itself. Although map from Yellowknife Bay and probably reflects Bay. The AF unit is defined by the Peace Vallis relations suggest that the BF unit represents distal lateral equivalents of bedrock studied by Cu- fan, which extends from the northwest rim of fan facies (rock types), these relations do not re- riosity, discussed further below. Gale crater (Fig. 1B); it may be the youngest quire that the Yellowknife Bay BF exposures The HP unit is exposed in the vicinity of component of a much more extensive alluvial be contemporaneous with the Peace Vallis lobe Bradbury Landing (14) and has lower values plain that flanks the crater wall (4). The Peace of the alluvial plain; they could represent an of thermal inertia. It is composed dominantly of Vallis fan probably formed by routing of eroded older component of the alluvial plain, separated loose materials, including clasts interpreted to be crater rim–derived rock fragments and any ma- from the youngest activity by an unconformity impact ejecta, but also displays erosional rem- terials that accumulated there after crater forma- (see below). Nevertheless, the close spatial asso- nants of underlying bedrock. Analyzed clasts tion, through Peace Vallis followed by downstream ciation implies a genetic link, with the BF unit have basaltic, alkalic (K- and Na-rich) compo- spreading and deceleration of flows and depo- having formed in a probably distal alluvial or sitions (15), and some clasts and soils contain sition of sediments in the downslope direction. even lacustrine (lake-related) setting, consistent crystals of probable feldspar (16). Windows The upslope part of the AF unit is marked by with earlier mapping (4). through this material in the vicinity of Bradbury smooth mottled surfaces punctuated by slope- Geologic units in the interior of Gale crater display a range of relative ages and preserve a substantial record of geologic activity since the crater formed close to the Noachian–Early Hesperian boundary ~3.7 billion years ago (6). The oldest units are probably of Early Hesperian age and likely include the bulk of the alluvial plain that extends downslope of the crater rim. The Peace Vallis alluvial lobe (map unit AF) ap- pears to represent the latest stage of transport and deposition on the alluvial plain. Our mapping shows that strata exposed at Yellowknife Bay are probably fan, distal fan, or downslope equivalents such as lacustrine depos- its. Despite the ambiguity in the relative age of these strata, the fluvial-lacustrine depositional con- text remains intact, and this is what motivated the MSL Science Team to explore these rocks.

Stratigraphy and Sedimentology of Yellowknife Bay Rocks Bradbury Landing occurs near the crest of a broad topographic high, several kilometers in width, informally called Bradbury Rise. To the north and east, the rise descends to rocks of the BF unit. After landing, the team-consensus de- cision was to traverse eastward and downhill to the HP-BF contact, as close as possible to the triple junction where the CS unit also inter- sects (Fig. 1B). Between Bradbury Landing and Yellowknife Bay, Curiosity crossed a straight- line distance of 445 m and descended 18 m in Fig. 2. Local context maps. (A) HiRISE image (PSP_010573_1755) of Bradbury Landing showing elevation. Most of this distance was spent tra- ’ Curiosity s position after landing and the path of eventual traverse to Yellowknife Bay, where the John Klein versing the clast-strewn HP unit, with its occa- and Cumberland holes were drilled. The area of this image is approximated by the circle and square sional outcrops of pebble conglomerate facies. symbols in Fig. 1B. Yellowknife Bay derives its name from Yellowknife crater, which is located within Beginning at the Bathurst Inlet outcrop, a suc- the broad expanse of the BF map unit. The BF unit probably represents the Yellowknife Bay formation at more regional scales. Cross section (A to A′) shows topography from Bradbury Landing to Yellowknife cession of strata was encountered with good Bay. (B) The inset shows a detailed geological map of the southwest part of Yellowknife Bay. Three maps exposure down to the floor of Yellowknife Bay units (HP, CS, BF) intersect just south of the Shaler outcrop. The BF map unit is subdivided into three (Fig. 2A). In ascending order, these strata are members (see text). Key outcrops are shown that provided information leading to the development of a informally named the Sheepbed (>1.5 m thick), stratigraphy shown in Fig. 4. Note that the contact between Sheepbed member and Gillespie Lake Gillespie Lake (~2.0 m thick), and Glenelg mem- member erodes to form a topographic step that is visible from orbit and can be traced for hundreds of bers (~1.7 m thick); the assemblage of members is meters around Yellowknife Bay. Therefore, it seems likely that much of the Yellowknife Bay topographic known as the Yellowknife Bay formation, and the depression was created by eolian erosion of the Sheepbed mudstone member (see text). See Fig. 4 for exposed section is ~5.2 m thick. (Fig. 3 and table mapped stratigraphic units. Locations of Rocknest (scoop location and rocky outcrops) and John Klein and S1). Strata are largely decimeter-scale in thick- Cumberland (drill) sampling locations are shown. ness and comprise a heterogeneous assemblage

www.sciencemag.org SCIENCE VOL 343 24 JANUARY 2014 1242777-3 Exploring Martian Habitability of mostly detrital sedimentary rocks of basaltic In general, the Yellowknife Bay formation A fluvial interpretation is further supported by bulk composition. Breaks in outcrop continuity (Fig. 4) is composed of fine-, medium-, and the tapered, likely sediment-filled cracks that due to accumulated loose materials on the slope coarse-grained sandstones of basaltic compo- are most easily accounted for by desiccation- (Fig. 3) introduce a degree of uncertainty in under- sition. The Shaler outcrop (~125 cm thick) consists induced shrinkage of finer-grained interbeds. standing lateral extent of bedding; the lowest strat- of a heterogeneous assemblage of interstratified These interbeds could represent either fluvial igraphic members (Sheepbed and Gillespie Lake) platy coarser-grained beds comprising sand- overbank deposits, slack water deposits in the are best exposed around the perimeter of Yellowknife stones and likely pebbly units, separated by troughs between bedform crests, or intercalated Bay (Fig. 2B). The Glenelg member contains a recessive, probably finer-grained intervals. Coarser- tongues of proximal lacustrine facies. diverse suite of facies formed of beds with de- grained intervals are characterized by distinct The Point Lake outcrop (~50 cm thick) is monstrably less lateral continuity, although most decimeter-scale trough cross-bedding (Fig. 5A). noteworthy for its pervasive centimeter-scale beds can be traced for at least tens of meters. Recessive intervals are marked by downward- voids (Fig. 5C). These voids could possibly point The basal Sheepbed and Gillespie Lake mem- tapering cracks that terminate upward at bedding to a volcanic origin, specifically a vesicular lava bers are characterized (19) by elemental com- planesandareprobablyfilledwithsedimentfrom flow (23). Alternatively, it is possible that this positions broadly consistent with a provenance the overlying bed. The rock termed “Rocknest-3” facies also represents a pebbly debris flow or that, on average, is similar to typical martian is very fine-grained, finely laminated sandstone gas-charged intrusive sedimentary sill (24, 25). upper crust (20). In addition, these rocks carry or siltstone with shrinkage cracks (Fig. 5B). Scattered blebs of light-toned material were eval- the signal of higher thermal inertia as seen from Volcanic or impact-related base-surge deposi- uated as calcium sulfate by ChemCam (19)and orbit. Subtle geochemical variations within and tion is excluded by the abundance of decimeter- appear to fill voids and narrow fractures. Thus, between these units mostly reflect minor varia- scale and compound cross-bedding, the absence it is possible that the rock could be a sandstone tions in provenance. A marked change in com- of stoss-side cross-stratification with preservation with leached nodular evaporites (26). position, best reflected by a sharp increase in only of bedform toesets, and variable cross-bed The Sheepbed and stratigraphically overly- K2O contents and K2O/Al2O3 ratios (Fig. 4), dip directions. Furthermore, there are hierarchi- ing Gillespie Lake members (Figs. 3 and 4) takes place higher in the section in the Glenelg cal scales of cross-stratification at the Shaler make up the exposed base of the Yellowknife member (Bathurst Inlet outcrop), suggesting a fun- outcrop, indicating migration of superimposed Bay formation, and their contact has eroded to damental change in the provenance to one dom- bedforms of different scales. These attributes are form a decimeter-scale distinctive topographic inated by more alkaline igneous rocks (16, 21). most consistent with a former fluvial environ- step observable in High Resolution Imaging This change could result from either propaga- ment characterized by bedload and suspended Science Experiment (HiRISE) orbiter image tion of the fluvial system headwaters into an load transport and variable, but net southeast data (Fig. 2A). The Gillespie Lake member con- area with more alkaline igneous rocks or a shift flow directions. However, some eolian bedload sists of amalgamated, sheetlike sandstones. Bed- to more locally derived volcanic debris, such as transport is suggested by intercalated beds with ding generally has a massive appearance, though reworked volcanic rocks or ash deposits (21). pinstripe lamination. Most paleocurrent data are poorly defined cross-bedding is present. The basal Major element compositions do not indicate sub- restricted to 90° of variability (fig. S1); however, bed of the Gillespie Lake member is formed of stantial chemical weathering in the sediment some scatter to almost 180°, which is consistent poorly sorted, angular to moderately well-rounded, source area, which implicates a frigid or arid with migration of three-dimensional (3D) super- medium to very coarse sand of bulk basaltic climate, a high-relief source area undergoing imposed bedforms (22). The inferred sediment composition (Fig. 5D). Some of the largest grains rapid erosion, or a combination of these fac- transport direction suggests derivation of sedi- have translucent luster. Very little residual po- tors (19, 21). ments from the direction of the Gale crater rim. rosity between grains is evident, which suggests

Fig. 3. Mastcam mosaic of Yellowknife Bay formation. View from the denote ChemCam or MAHLI only. Both scale bars are 50 cm long with 10-cm- base of the exposed section up through Sheepbed, Gillespie Lake, and basal spaced white and black sections. The lower scale bar is ~8 m away from Cu- Glenelg members. Locations of drill holes and APXS measurements are shown. riosity; the upper scale bar is ~30 m away. The 111-image mosaic was acquired High density of APXS measurements between Snake River and Ungava targets onsol137,byM-34sequence818.Thisfigure shows only a small portion of delineates the route of the Selwyn reference section. Note the “snake” feature the full mosaic. The foothills of Mt. Sharp are visible in the distance, upper cutting up through the section and terminating in a small anticline. White dots left,southwestofcameraposition.Formosaic image identification numbers represent combined APXS, MAHLI, and ChemCam measurements; gray dots (IDs), see (31).

1242777-4 24 JANUARY 2014 VOL 343 SCIENCE www.sciencemag.org RESEARCH ARTICLE a high degree of cementation; what porosity does and the absence of well-developed channel bodies voke superposition and, therefore, a younger age exist appears as sparse, millimeter-scale irregular may indicate deposition as distal fan lobes, as for the conglomerate than the Yellowknife Bay vugs (void spaces). This basal bed has a sharp, documented in examples from ancient alluvial formation. However, regional mapping allows erosional base that shows centimeter-scale scour- fans on Earth (27). for, but does not require, an alternative possi- ing into the underlying Sheepbed member mud- The pebble conglomerate outcrops distributed bility in which flat-lying Yellowknife Bay rocks stones and appears to show lateral variation in across the HP map unit have an uncertain strati- unconformably onlap older rocks that underlie its thickness. The poor sorting, massive bedding, graphic relation with respect to the Yellowknife the HP map unit; in this interpretation, the con- and variability of grain shape in the Gillespie Bay formation. On one hand, the conglomerate glomerate is relatively older. Unfortunately, out- Lake sandstones are most consistent with fluvial exposures occur at a topographically high level crop between Yellowknife Bay rocks and the transport and deposition. The sheetlike geometry above flat-lying rocks of the Yellowknife Bay nearest conglomerate is very poor, and it is not of the beds, extending over hundreds of meters, formation; the simplest interpretation would in- possible to observe a contact. Therefore, both

Fig. 4. Stratigraphy of the Yellowknife Bay formation plotted with ele- of stratification styles; primary textures; and diagenetic textures including mental and oxide ratios measured by the Curiosity APXS instrument. The nodules, hollow nodules, raised ridges, sulfate-filled fractures, and vugs. Yellowknife Bay section was split into three members distinguished by lith- APXS analyses are plotted as ratios to highlight compositional trends through- ological properties (facies, textures), chemical composition, and attributes out the Yellowknife Bay formation. Average Mars crustal composition for each observed from orbit by HiRISE. The stratigraphic column highlights differ- ratio (colored vertical lines) (20) is plotted for comparison to Curiosity APXS ences in grain size between the Yellowknife Bay members and the presence measurements.

www.sciencemag.org SCIENCE VOL 343 24 JANUARY 2014 1242777-5 Exploring Martian Habitability options are feasible, but in either case, these file of the Sheepbed–Gillespie Lake contact shows X-ray Spectrometer (APXS), and Mars Hand rocks all belong to the same genetically related that the Sheepbed member probably extends Lens Imager (MAHLI) measurements was con- suite of fluvial-lacustrine sediments derived from around the full width of Yellowknife Bay and ducted across the Sheepbed member for the the crater rim. covers a minimum area of ~4 km2. However, we purpose of providing context for the first sam- presume the Sheepbed member extends beneath ples obtained by drilling (Figs. 3 and 4). Ob- Sheepbed Mudstone Member the overlying rocks of the Gillepsie Lake mem- servations revealed a patchy distribution of At a minimum, the Sheepbed member extends ber; therefore, these are highly conservative mini- diagenetic (postdepositional) features, includ- laterally across the width of the outcrop belt mum estimates. A reference section localized near ing nodules and veins, as discussed below. Min- traversed by the rover: ~60 m. The Sheepbed the Selwyn target (Fig. 3) yields a thickness of eralogic data show that the Sheepbed member member closely coincides with the BF map unit 1.5 m for the member. This is a minimum thick- has a substantial quantity of saponitic smectite and from orbit is highlighted by its light tone ness because its base is not exposed. clay minerals (~20%), probably formed by aque- and extensive meter- to decameter-scale fracture A detailed traverse (Selwyn reference sec- ous alteration of olivine (28). Geochemical re- network. Tracing of the distinctive erosional pro- tion) using Mastcam, ChemCam, Alpha Particle sults (19) show that both Sheepbed and Gillespie

Fig. 5. Sedimentary rocks of the Yellowknife Bay formation. (A) Shaler outcrop shows a variety of facies, including com- pound cross-bedding. The image was white-balanced and acquired by Mastcam- 100 on sol 120. Image ID is 0120MR00075201302- 00690E01_DXXX. (B)The Rocknest-3 rock, which rep- resents the middle part of the Glenelg member, is fine- grained with planar lam- ination, shrinkage cracks, and millimeter-scale voids (well developed in the lower third section of the rock). This image is a subset of a mosaic of images that were white-balanced and acquired with Mastcam-100 on sol 59. Image IDs, as part of SEQID mcam00270, are 0059MR00027000001031- 41E01_DXXX, 0059MR000- 2700010103142E01_DXXX, 0059MR00027000201031- 43E02_DXXX, and 0059MR- 0002700030103144E01_ DXXX. (C) Vuggy, medium- grained texture in the Point Lake outcrop (“Measles Point” target), middle of the Point Lake member. Note the thin fractures filled with light- toned material, interpreted as sulfate. ChemCam data show that white blebs, such as the one in the center, are composed of calcium sulfate. This focus merge product was acquired by MAHLI on sol 303, ID spaces (gray arrow), and sulfate-filled voids can be seen, and only those 0303MH0002900000103786R00. (D) Medium- to coarse-grained sandstone voids connected by hairline fractures (white arrow) have been filled at the Gillespie Lake outcrop (see Fig. 2B), basal bed of the Gillespie Lake with sulfates. Note the surface scoring by brush bristles, indicating member. This bed is composed of poorly sorted, moderately well-rounded, softness of mudstone. The image was acquired by MAHLI on sol 169, ID dark to light grains, some of which have glassy luster (lower left arrow). A 0169MH0002080020102218C00. (F) John Klein drill hole (see Fig. 3) small amount of vuggy porosity is evident (upper right arrow), but otherwise reveals gray-colored cuttings, rock powder, and interior wall. Note the the rock appears well cemented. The image was acquired by MAHLI on sol homogeneous, fine grain size of mudstone and the irregular network of 132, ID 0132MH0001580010101221E01. (E)MudstoneinSheepbedmem- sulfate-filled hairline fractures. An array of eight ChamCam shot points ber, “Wernecke target” (see Fig. 3). The rock has been brushed, revealing can be seen. The diameter of the hole is 1.6 cm. The image was acquired homogeneous, fine-grained texture. Outlines of nodules (black arrow), void by MAHLI on sol 270, ID 0270MH0002540050102794C00.

1242777-6 24 JANUARY 2014 VOL 343 SCIENCE www.sciencemag.org RESEARCH ARTICLE sediments were derived from a provenance with all, the primary texture and grain size has a very the Sheepbed mudstone, which is closer to 1% a composition similar to average martian crust and uniform, massive appearance; it is well known, SO3 (19). Although it seems conceptually pos- that experienced negligible chemical weathering however, that lamination can be difficult to de- sible to deposit relatively thick intervals of dust before deposition at Yellowknife Bay. Relatively tect in many mudstones (29). (35, 36), this might not be expected on an al- subtle differences exist between the lower and The very fine grain size, its apparently uni- luvial fan that is intermittently active enough to upper parts of the Sheepbed, such that the top form distribution through the member, and the result in the substantial amounts of coarser-grained exhibits higher Al2O3/TiO2 and Ni, consistent substantial lateral extent of centimeter-scale in- sediment accumulation seen in the Yellowknife with a slight provenance change. A further change terbeds are all consistent with deposition of the Bay formation. Furthermore, estimates of dust in trace elements (higher Cr, lower Ni and Ge), primary grains due to settling from suspension, accumulation rates are very low [on the order which occurs across the Sheepbed–Gillespie Lake either from the atmosphere or a water column. of 10 cm per million years (37)], as compared contact, could result from either provenance dif- The former might involve volcanic ash, eolian with fluvial or even lacustrine sediment accu- ferences or mineral sorting (e.g., chromite) associated dust, or very distal impact fallout. Because the mulation rates, which are orders of magnitude with grain-size differences. Both APXS mea- composition of the Sheepbed mudstone is ba- higher (38). surements and high–spatial resolution ChemCam saltic and we cannot ascertain the very finest Sheepbed compositions likely discount an targeted analyses provide evidence that crosscut- grain sizes, it is hard to discount an airfall ori- origin as pure volcanic ash. Apart from variation ting veins and fractures are composed of Ca- gin. However, mapping relations, regional con- imposed by sulfate-rich diagenetic features, the sulfate (see below). text, and local sedimentologic constraints all Sheepbed is geochemically uniform with an ad- The highest-resolution MAHLI images were point to a distal alluvial fan or proximal lacus- justed volatile-free composition close to aver- acquired on sol 150 (Ekwir) and sol 169 (Wernecke) trine setting. In this setting, sediment-laden fluvial age martian crust (19). Individual volcanic ash after brushing of the outcrop (Fig. 5E). These flows can deposit extensive fine-grained depos- eruptions might be expected to display some images reveal a very fine grain size based on the its, either as overbank flood deposits or at their variation in composition, and it would be co- outlines of darker grains, and nearly all discern- downstream termination where they enter a body incidental for individual eruptions to all have ible grains are finer than 50 mm. By definition, a of standing water, such as a lake. Here, the de- a composition essentially the same as the av- mudstone is a fine-grained sedimentary rock crease in flow turbulence as flows thin overbank erage crust. If the >1.5-m-thick Sheepbed mem- composed of 50% or more particles smaller than and spread when leaving a channel, or encounter ber represented a single eruption, this would 62 mm(29). Although we cannot directly resolve the still water of a lake, leads to deposition of require a local source that is not observed in the and observe these <50-mm grains, we do know suspended fines. We favor the lake interpretation vicinity of Gale crater [probably within a few that ~20% are clays. It seems likely that the entire because of the lateral extent and thickness of the tens of kilometers for 1 m of thickness (39)]. Sheepbed member is finer than most facies ob- Sheepbed member, the absence of fluvial chan- served by Curiosity during the traverse down to nel bodies laterally adjacent to mudstone beds, Diagenesis of the Sheepbed Mudstone Yellowknife Bay. Drilling of the outcrop produces and the fact that the mudstone underlies the A diverse set of diagenetic features that postdate gray powder and cuttings that are substantially Gillespie Lake member fluvial sheet sandstones deposition is observed in the Sheepbed mudstone different in color from the modern surface (Fig. over a large areal extent. Given our limited ob- member. Early diagenetic features include nod- 5F). Inspection of the drill hole reveals a gray servations, it is difficult to exclude the hypoth- ules, hollow nodules, and raised ridges (Figs. 6, A rock that is fine-grained and homogeneous in esis that this lake may have been seasonally dry to C). These features are crosscut by later frac- texture, with the exception of light-toned frac- (playa), or whether it responded to longer-term tures, filled with light-toned cements (Fig. 6D) tures filled with minerals. wet-dry climatic fluctuations. However, the com- shown to be variably hydrated sulfate minerals Independent evidence for the fine grain size positional data favoring low salinity for the lake basedonChemCam,APXS,Mastcam,andChem- of the Sheepbed mudstone comes from infer- waters (19) and specific diagenetic textures istry and Mineralogy instrument (CheMin) data. ences of the physical responses of brushing and discussed below (cement-filled synaeresis cracks, Additionally, a dike-like feature known as the measurements made during drilling. The rotary rather than sediment desiccation cracks) suggest “snake” obliquely crosscuts the Sheepbed–Gillespie motion of bristles during brushing substantially that it was perennially wet for the stratigraphic Lake contact and extends for several meters (Fig. 3). scored the surface of the outcrop in a fashion interval we examined. Concentrations of nodules and hollow nod- that shows deflections (Fig. 5E) around appar- Strata dominated by fluvial sediments that ules have patchy distributions and pass laterally ently harder, embedded dark nodules, suggesting predate the advent of terrestrial vegetation on (Fig. 7) into mudstones that have raised ridges. that the mudstone matrix is softer. Comparison Earth are generally devoid of fines, except as Nodules are expressed as millimeter-scale pro- with terrestrial analog data suggests weak to me- occasional mud drapes that are preserved be- trusions of the outcrop with 3D differential relief dium hardness for the mudstone (30). Additional tween sets of cross-stratified dune or bar depos- suggesting crudely spherical geometry (Fig. estimates of rock hardness are provided by com- its (32, 33). Sweeping of very broad and shallow 6A). The mean diameter of 4501 nodules is parison of rover engineering testbed data derived channels that produce sheet flood deposits result 1.2 mm, with minimum and maximum diame- from terrestrial analogs with drill hardness data in reworking of thin floodplain deposits that ters of 0.4 and 8.2 mm, respectively. Brushing from the Sheepbed member (31). These data sets tend to not be preserved in the stratigraphic shows that nodules are more resistant to brush- suggest that the Sheepbed rock is comparable to record. In addition, eolian reworking between induced etching and, thus, harder than the sur- fluvial-to-lacustrine siltstones and mudstones flow events, in the absence of terrestrial vegeta- rounding mudstone (Fig. 5E). Mastcam, MAHLI, collected from the ~10-million-year-old Ridge tion to anchor fines, is also regarded as an ef- and ChemCam Remote Micro-Imager (RMI) Basin of southern California (buried to depths of ficient mechanism to sweep fines away from data show that nodules are dispersed from one several kilometers) but not as hard as feldspathic sites of deposition. Thus, a shallow lacustrine another, cluster in patches (Fig. 7), do not form sandstones derived from the same basin or as environment, which sequesters fine sediments their own beds, and show no evidence of size soft as pure commercially available kaolinite. over long time scales, is considered to be the grading, despite a broad range in diameters. These Sheepbed mudstones contain a few very thin most likely interpretation for the Sheepbed constraints suggest that the nodules are of con- intercalated beds that resist erosion (Fig. 4). mudstone. cretionary origin rather than volcanic or impact These interbeds have not been studied in detail, The grain size and composition may be in- lapilli (40, 41). Furthermore, no associated out- but Mastcam data show them to be 1 to 2 cm consistent with typical Mars dust, which is even sized clasts that might be interpreted as potential thick and with substantial lateral continuity, as finer in grain size and characterized by SO3 bombs, or any breccias, are observed. Compar- some can be traced for more than 100 m. Over- contents approaching 7% (20, 34), rather than ison of John Klein (fewer nodules and hollow

www.sciencemag.org SCIENCE VOL 343 24 JANUARY 2014 1242777-7 Exploring Martian Habitability nodules) and Cumberland (more nodules and ules are similar in size to nodules with mean which creates the hollow observed in outcrop: In hollow nodules) compositional data indicates diameters of 1.2 mm, as well as minimum and the first hypothesis, the void space would repre- that nodule and hollow-nodule rim growth may maximum diameters of 0.6 and 5.6 mm, respec- sent a vanished mineral, whose higher solubility have involved iron-bearing compounds such as the tively (n = 1248 hollow nodules). We note that allowed easy dissolution during subsequent dia- mineral akaganeite or even clay minerals (19, 28). the upper size limit of hollow nodules is consid- genetic fluid percolation events. Such minerals Hollow nodules are millimeter-scale circular rims erably smaller than that for nodules. The interior might have grown displacively in the sediment with hollow centers (Fig. 6B). Three-dimensional mean diameter of the void space defining the hol- [for example, in evaporites (26)], thereby elimi- exposures sometimes reveal pedestal-like struc- low is 0.7 mm (n = 45). nating the need for the void to have been primary. tures with a circular rim surrounding a hollow, Hollow nodules are regarded as void spaces However, no examples of filled hollow nodules bowl-like depression. The outcrop therefore has preserved within the Sheepbed mudstone. How- were observed, with one very specific exception: a locally pitted appearance, suggesting erosional ever, the hollow nodules are distinctly different Where hairline fractures filled with light-toned exposure of approximately spherical voids that had from millimeter-scale diagenetic vugs, such as sulfate minerals intersect hollow nodules, the hol- resistant rims, similar to the nodules. Resistance in the Gillespie Lake member or Burns forma- low nodules also are filled with sulfates. Other- to weathering by the rims of the hollow nodules is tion of Meridiani Planum (40), in that the hol- wise, hollow nodules are devoid of any remnant indicated by the presence of lineations and “wind low nodules are associated with a ubiquitous of an earlier mineral. The sulfate-filled fractures tails” carved into the outcrop, presumably by circumscribed rim. Spherical voids of this type represent a later phase of diagenesis (see below), eolian saltation abrasion. Development of pedes- have not been observed previously on Mars and and therefore, the sulfate minerals are not con- tals likely records lowering of the outcrop surface perhaps relate to their ability to be preserved— sidered part of the primary environment. by prolonged eolian abrasion. Mastcam, MAHLI, in this case, within fine mudstone sediments. Two This allows for a second hypothesis in which and ChemCam RMI data show that hollow nod- options are possible to explain the void space, the hollow nodules represent primary voids,

Fig.6.Diagenetic fea- tures of the Sheepbed member. (A) The Sheep- bed target shows that millimeter-scale nodules, interpreted as concretions, are locally abundant. Note their dispersed distribu- tion. The outcrop is crosscut by sulfate-filled fractures. This image is a subset of a mosaic acquired by Mastcam34 on sol 192. Image IDs are 0192ML- 0010170240105705E01_ DXXX and 0192ML00101- 70070105688E01_DXXX. The image has been white- balanced. (B) Hollow nod- ules at the Bonnett Plume target (see Fig. 3) are de- fined by circular cross- sections through voids with raised rims (white ar- row). Voids are filled with sulfate minerals where in- tersected by hairline frac- tures (black arrow) but are otherwise unfilled. The im- age was white-balanced and acquired by Mastcam- 100 on sol 159. Image ID is 0159MR0008640050- 201360E01_DXXX. (C) Raised ridges near the McGrath target have spindle- shaped terminations (see Fig. 3). Note the broadly varying strikes with dips ranging from vertical to 45°. Ridges most commonly mineralizing fluids. Note the abundant millimeter-scale nodules. The frac- have parallel sides, but they also have one or two additional infilling bands. All ture fills are mostly flush with the bedrock surface, which contrasts with the contour the local topography of the fracture margin and are interpreted as nodules that weather in relief. This image is a ChemCam RMI mosaic that has infilling cements. The image was white-balanced and acquired by Mastcam-100 been pan-sharpened using white-balanced Mastcam-100 color data. Images on sol 164. Image ID is 0164MR0008850000201589E01_DXXX. (D) Sulfate- used in this composite are CR0_408676341EDR_F0051398CCAM01126M1, filled fractures at the Sheepbed target (see Fig. 3) are consistent with brittle CR0_408675268EDR_F0051398CCAM01126M1, and Mastcam-100 ID deformation of lithified mudstone, followed by precipitation of sulfate from 0126MR0007800000200786E01_DXXX.

1242777-8 24 JANUARY 2014 VOL 343 SCIENCE www.sciencemag.org RESEARCH ARTICLE such as gas bubbles formed during or soon after on Mars. Terrestrial sediments are pervaded by shaped, pointed terminations (Fig. 6C). Most deposition. Several mechanisms could gener- microbes, which produce a variety of gases that have subvertical orientation but can dip more ate gas during or after sedimentation, including become trapped as bubbles where lithification is gently, down to 45° or less, and their strike trends diagenetic dissolution and precipitation reactions early and rapid (43). However, it should be clear in all directions (Fig. 6C). Intersections of 90° that might produce gas as a by-product. These enough that this mechanism can only be in- between ridges are common. A distinctive at- might include hydrogen gas produced during voked as a serious possibility after all other abiotic tribute is that most ridges show internal banding, authigenic clay mineral formation via olivine hypotheses have been discounted. This is not the which contours the trend of the ridge, following “saponitization” (28); ultraviolet (UV) photo- case for our current data set, and we include this closely all local deviations, including right-angle oxidation of reduced minerals, such as siderite, only for the sake of completeness. bends. Two parallel bands are most common, to produce authigenic iron oxide (e.g., magnetite) Raised ridges (44) are narrow, curvilinear although three to four are also observed (Fig. 6C). and hydrogen gas (42); pressure-induced release ridges that weather in raised relief, and, as men- Raised ridges are most simply interpreted as early of dissolved gas from sedimentary pore fluids; tioned above, they substitute laterally for nod- diagenetic cracks that formed in consolidated and processes that reflect distinct Mars environ- ules and hollow nodules (Fig. 7). The raised or even partly lithified mudstone, then filled ments, such as boiling of water in the presence ridges are restricted to the Sheepbed mudstone, with one or more generations of cement that of critically low atmospheric water vapor pres- but one poorly constrained example was noted isopachously encrust the margins of the cracks. sure. Data are insufficient to distinguish among at a higher stratigraphic level (observed only This infilling material has greater resistance to these possibilities. However, the last of these from a distance; exact stratigraphic position and eolian erosion than the mudstone, resulting in can be discounted by the presence of thin inter- facies are uncertain). Measurement of 1619 raised its expression as ridges. ChemCam and APXS beds within the Sheepbed mudstone, which sug- ridges in the Sheepbed suggests that they extend analyses have identified elevated Mg, Fe, Si, gests that wholesale churning of the mudstone at least several centimeters in length and have a Cl, Li, and Br in the raised ridges, but insuf- did not occur. A final possibility considers the mean width of 2.7 mm, with maximum widths ficient evidence exists to confidently identify potential role of microbes, if life had ever evolved of up to 5.8 mm. Raised ridges have spindle- specific mineral phases (19). A strong corre- lation exists between Mg, Fe, and Cl, possibly indicating that a chloride phase may be present, Fig. 7. Map of diagenet- but if so, it can only explain a small amount of ic features showing spa- the Mg and Fe enrichment. Alternatively, the tial relations between correlation could result from a phase other than fabric elements. The rock a chloride that contains substantial Cl. For ex- surface coincides with a ample, CheMin data suggest the presence of bedding plane. Note that akaganeite (28). nodules and hollow nod- Taken alone, the development of very early ules pass laterally into raised cracks in fluvial-to-lacustrine mudstones might ridges, suggesting varia- be most simply explained by desiccation of wet tions in lithologic or dia- sediment: Generally, desiccation cracks have po- genetic fluid properties lygonal geometry. Therefore, the nonpolygo- (see text). (A) A subset of the Mastcam-100 work- nal geometry of the cracks is unexpected, as are space mosaic acquired by the isopachous, banded materials that fill in the the Curiosity on sol 164. fractures. Furthermore, the abundance of raised This perspective mosaic is ridges is anticorrelated with the abundance of white-balanced and was nodules and hollow nodules, suggesting that all used as a base map. (B) three phenomena are genetically related, which The sol 164 workspace mo- is not observed for desiccation cracks on Earth. saic annotated to show An alternative mechanism proposes that syn- the distribution of early depositional pore fluids (perhaps mixing with diagenetic features (nod- surface runoff waters) may have been important ules, hollow nodules, raised in driving diagenetic reactions that triggered ridges) and later diage- early mineral precipitation, gas production, and netic features (thick to thin volume contraction close to the sediment-water sulfate-filled fractures). interface. On Earth, a related but poorly under- Erosion-resistant raised ridges stood process known as “synaeresis” is thought were outlined individually to relate to clay mineral flocculation during and generally exhibit par- mixing of saline and less saline waters, resulting allel sets of isopachous in differential contraction of the substrate (45). linings. Thin (<5 mm in Seismic shaking, perhaps induced by impacts on width) sulfate veins have Mars, may help facilitate this process (46). The relatively constant thick- ness, whereas thicker sulfate result is spindle-shaped cracks, a few centime- veins have more variable ters long, that look very much like the Sheepbed thickness. Both sets of veins raised ridges. In most cases on Earth, the cracks crosscut other diagenetic are filled by other sediments; however, in one features.Hollownodulesarecircularfeatures(~1mmindiameter)thatstandoutinraisedrelieffrom very particular instance known as “molar tooth the outcrop surface and are distinguished from nodules by the presence of a small depression in the structure,” spindle-shaped cracks and related sub- center of the feature. Because hollow nodule diameters are often at or below the resolution limits of spherical voids are filled with mineral (calcite) this mosaic (confirmed by MAHLI observations), hollow nodules are probably underrepresented relative cement. The origin of subaqueous cracks on to nodules. This image is a subset of the Mastcam-100 workspace mosaic from sol 164. For mosaic Earth is poorly understood, though the involve- image IDs, see (31). ment of gas in generating the cracks and voids

www.sciencemag.org SCIENCE VOL 343 24 JANUARY 2014 1242777-9 Exploring Martian Habitability represented by molar tooth structure is likely sedimentary injection (24, 25). This suggests a larger grains, or perhaps clusters of grains, were (43, 47). As with the hollow nodules, a number of sediment accumulation rate for the Yellowknife leached during fluid circulation. Alternatively, it processes could generate gases in martian sedi- Bay formation (or sediments below it) that was is possible that these vugs represent intraclasts of ments. In general, the distribution of subaqueous high enough so that pore fluid pressures built up mudstone that degraded by weathering. Another shrinkage cracks reflects differential strength of to the point where expulsion along discrete zones possibility is dissolution of preferentially more the substrate materials, which may similarly be of weakness, rather than along grain boundaries, soluble early diagenetic minerals; however, the represented in the Sheepbed member by the spa- was the preferred pathway for escape of compact- irregular shape of the vugs is not consistent with tial segregation of the nodules and hollow nod- ing sediments and interstitial pore fluids. As the dissolved evaporite crystals. ules versus raised ridges. porosity and permeability of the sediment was The diagenesis of the Sheepbed member A distinct feature within lower Yellowknife reduced via compaction and cementation, the strata suggests a complex aqueous history involving at Bay strata is a dike-like, dark-toned rock unit may have ruptured, allowing fluids to escape and least two distinctly different water masses. After referred to as the “snake,” named after the target form sedimentary dikes and sills. Finally, this pro- emplacement in distal alluvial-to-lacustrine envi- Snake River (see Fig. 3). This feature extends cess could have been aided by impact-induced ronments, pore spaces in primary sediments from the lowest stratigraphic level of the ex- seismic activity (50). were occluded by cements, and early diagenetic posed Sheepbed member, crosscuts the Sheepbed– Fractures filled with light-toned sulfate ce- hollow nodule voids (gas bubbles?) may have Gillespie Lake contact, and tapers out within the ments also are regarded to have formed later been formed, along with concretions. At approx- middle of the overlying Gillespie Lake member in the diagenetic sequence because they cut the imately the same time, nonpolygonal fissures in (Figs. 3 and 4). The snake is 6 to 10 cm wide entire Yellowknife Bay formation. In contrast to the lithifying sediment opened up and became and can be traced across bedding for ~5 m, the raised ridges, the sulfate-filled fractures are lined with void-filling isopachous rim cements. cutting up through almost 1 m of stratigraphic flush, even slightly depressed, with respect to ad- All of these events, with the exception of hollow section. MAHLI images of unbrushed parts of jacent unfractured outcrop (Figs. 6 and 7). Sulfate- nodule void formation, must have occluded much the snake suggest that it has a fine-grained clastic filled fractures have highly diverse orientations of the primary porosity. This episode of diagen- texture, but with larger fragments present sug- spanning from vertical to subhorizontal; range esis did not substantially change the bulk rock gestive of intraclasts (fig. S2). Elemental data inlengthfromlessthan1cmtomorethan30cm; chemistry of these units relative to average show it to be similar in major elements but dis- and range in width from “hairline” (submilli- martian basalt and are thus considered to have tinct from the Sheepbed member in trace ele- meter) to 8 mm, with a mean width of 2.3 mm involved isochemical alteration (19). This sug- ment abundances (19). The snake can be traced (Figs. 5E and 6D). MAHLI images of the John gests that clay mineral and magnetite formation to its stratigraphically highest position, where Klein drill hole show that sulfate-filled fractures happened largely in situ, as a result of alteration it terminates in a small anticline. At this terminus, of hairline width have complex orientations with of olivine to form authigenic saponitic smectite thicker, massive beds extend laterally away from horizontal as well as vertical trends (Fig. 5F). (19, 28). In addition to pore-filling cement, it is the core of the anticline. The snake has several Where these hairline-width fractures intersect possible that these minerals also precipitated as possible interpretations, including a narrow ig- hollow nodule void spaces (Fig. 5E), the latter concretions, as well as isopachous rim cements neous dike of basaltic composition; a large-scale are also filled with hydrated and anhydrous cal- within raised ridge fractures. After lithification, crack that is filled from above, perhaps asso- cium sulfate minerals, as determined by CheMin, fracture porosity was created that was then filled ciated with subaerial exposure or ice wedging; ChemCam, and APXS data (19, 28). At a larger with a distinctly different fluid that resulted in or a sedimentary intrusion associated with com- scale, the observation of fracture-filling hydrated precipitation of sulfate cements (19, 28). The paction and dewatering of the mudstone or sub- sulfate mineralogy can be extended through Mastcam absence of elevated sulfate abundances in the jacent strata. observations. Calibrated Mastcam spectra show Sheepbed mudstone, except for within these frac- The fine-grained, basaltic composition of the evidence for hydration associated with most but tures, indicates that cementation and permeability snake makes it hard to exclude an igneous in- not all of the fracture fills, including limited hy- reduction of the mudstone was extensive enough trusion with basaltic composition. The feature is dration in hairline fractures that penetrate and so that sulfate-rich fluids percolating through so thin that grain-size variations consistent with fill the void space of hollow nodules (28). fractures and hairline fractures did not invade chilled margins would be unexpected at scales To form, the sulfate-filled fractures first re- the mudstone or sandstone matrix (19, 28). The within MAHLI’s limit of resolution; the entire quire lithification of the Sheepbed and Gillespie first diagenetic fluid, possibly derived from the dike would be expected to be very fine grained. Lake members to allow applied stress to cause waters that transported sediments, was dominated Filling from above is discounted due to the pres- brittle strain of the bedrock (51). Once open, the by dissolved silicate and iron oxide minerals. The ence of the small anticline, which suggests in- fracture networks then act as conduits for fluid second, subsequent fluid was dominated by jection of fluidized material from below rather transport, resulting in precipitation of cementing sulfates and perhaps sampled a broader, basin- than settling of sediment from above. In ad- minerals within these initially low permeability wide hydrologic system. CheMin and Sample dition, no evidence of layering, suggestive of sedimentary rocks. The fracture networks were Analysis at Mars instrument (SAM) data addi- downward percolation of sediments, is observed possibly created by hydraulic fracturing due to tionally point to a substantial x-ray amorphous within the feature. Furthermore, no other crack- high fluid pressures that would have been ob- fraction of material that could have been in- like features are observed to extend downward tained during burial of Yellowknife Bay strata volved in diagenesis, as well as potential sulfide from a common stratigraphic surface. The third (51). The variation in hydration state of vein- minerals (28, 52). possibility—intrusion of plastically deformed sed- filling sulfate mineralogy between gypsum, imentary materials—occurs in some sedimenta- bassinite, and anhydrite could reflect primary Sheepbed Mudstone: Record of ry environments with high sedimentation rates variations in environmental conditions during Habitable Environment (24, 25, 48, 49). Upward and lateral injection of deposition or, alternatively, diagenesis associ- The regional geologic context, sedimentologic compacting muds and sands is most commonly ated with burial and/or desiccation during sub- framework, and geochemistry and mineralogy expressed as dikes and sills that both crosscut sequent erosion and exposure (28). of the Sheepbed mudstone all point to an ancient strata and intrude along bedding planes. The A final observation regarding later diagenesis environment that would have been habitable similarity of both composition and texture to concerns the Gillespie Lake sandstone member, to a broad range of prokaryotic microorga- that of the Sheepbed mudstone, the presence which shows evidence for possible secondary nisms. The relatively neutral pH indicated by of possible intraclasts, and upward deflection porosity. Irregular vugs larger than the diameter clay mineral stability (28); high water activity of strata at its terminus are all consistent with of the average grain size (Fig. 5D) suggest that indicated by low salinity for primary and early

1242777-10 24 JANUARY 2014 VOL 343 SCIENCE www.sciencemag.org RESEARCH ARTICLE diagenetic environments (19); the variable redox proach involves identification of the chemical and SAM data. Further, observation of key in- states indicated by presence of magnetite, sul- and physical environmental requirements that organic volatiles such as H2O, H2S, SO2,H2, fate, and sulfide minerals and compounds (28, 52); would support prokaryotic microbes, if similarly and CO2, generated during SAM pyrolysis (52), and the apparent long duration of the aqueous simple forms of life had ever evolved on Mars. suggests a wider range of redox conditions for environment(s) all support this interpretation. Bacteria and archaea represent ancient, abun- the surface of ancient Mars than previously recog- The Sheepbed mudstone closely coincides dant, and metabolically diverse domains of life nized based on Viking, Phoenix, and Mars Ex- with the BF map unit (Fig. 1), suggesting that on Earth and are able to withstand a wide range ploration Rover data (71–74). For example, SAM this rock unit may have extended beyond the of environmental conditions (58–61). They can analysis of the John Klein sample shows higher confines of Yellowknife Bay and minimally cov- engage in a multitude of autotrophic and het- H2S/SO2 average ratios than were found for ered an area of ~30 km2, not including buried or erotrophic pathways and use a variety of electron the Rocknest soil sample: 0.06 for John Klein eroded equivalents. Repeated flows of water (17) donor-acceptor pairs for acquisition of carbon (52) versus 0.01 for Rocknest (75). These data and the likely presence of a body of standing and energy (62–65). We can envisage many com- suggest a higher proportion of reduced sulfur water point to a sustained habitable environment. binations of terrestrial microbes that would be relative to oxidized sulfur pointing to a plausi- The durations of the primary, early diagenetic, and suited to form a martian biosphere founded on ble redox couple for prokaryotic respiration. later diagenetic aqueous environments were sub- chemolithoautotrophy at Yellowknife Bay. Iron also presents intriguing possibilities for stantial enough to create geologic and geochem- As a first step, it is useful to assess the key redox coupling. CheMin analyses indicate the ical fingerprints of prolonged fluvio-lacustrine nutrients required to support terrestrial bacteria presence of a substantial quantity of likely au- and diagenetic processes markedly similar to and archaea (66). The defining chemistry is in- thigenic magnetite (~7% of crystalline compo- those that formed commonly during early Earth organic: elemental inventory, liquid water, redox nents) in Sheepbed mudstones (28). The presence history (32, 53, 54). potential (Eh), pH, redox couples, and a diversity of an authigenic, partially reduced Fe-oxide phase A rough estimate of the minimum duration of chemical compounds. Organic molecules are in a martian sedimentary rock represents a no- of the lacustrine environment is provided by the not required environmental constituents because table departure from previous secondary Fe- minimum thickness of the Sheepbed member. they can be synthesized de novo through the prim- mineral detections on Mars, where ferric iron has Given 1.5 m, applying a mean sediment accu- itive autotrophic pathways, such as methano- proven to be the dominant redox state (76–78). mulation rate for lacustrine strata of 1 m per genesis and acetogenesis (67). We have measured The magnetite measured by CheMin does show 1000 years (38) yields a duration of 1500 years. H, O, S, C, N, and P in minerals and other com- evidence for oxidation of some of its ferrous Lacustrine sediment accumulation rates vary by pounds with the APXS, ChemCam, CheMin, iron to form the ferric defect-spinel maghem- several orders of magnitude (38); thus, the hy- and SAM instruments, as well as important me- ite (28). The gray color of the freshly exposed pothesized lake could represent hundreds to tals such as Fe, Mg, and Mn (19, 28, 52). All Sheepbed mudstone (Fig. 5F) contrasts sharply tens of thousands of years. If the aqueous envi- but P are observed in SAM evolved gas analysis with the reddish color of the freshly exposed, highly ronments represented by overlying strata are (EGA) data as volatiles, and all but P and N are oxidized, hematite-bearing rocks of Meridiani considered, such as the Gillespie Lake and known to be part of mineral structures observed Planum (79), emphasizing the lower oxidation Glenelg members, then this duration increases. in CheMin data. We do not have the capability state of iron in the Sheepbed mudstone. There Indeed, the regional geologic context suggests to determine whether P would have been bio- are a number of potential mechanisms for the that the Yellowknife Bay formation may be just available,andCheMinobservednohighlysoluble production of authigenic magnetite, all of which a small part of a thick series of fluvio-lacustrine P-bearing minerals. On the other hand, CheMin are promoted by waters with relatively low Eh rocks that may exist in the subsurface for hun- data indicate substantial breakdown of olivine in and neutral to alkaline pH (80). Under such con- dreds of meters. Moreover, it is possible that the mudstone at Cumberland (28), and P can be ditions, aqueous Fe2+-hydroxide species can pre- hundreds of meters of fluvial-lacustrine strata a trace-to-minor element in that mineral (68), cipitate and ultimately transform to magnetite. have been eroded. The sulfate-filled fractures are suggesting potential P mobility during diagene- Perhaps not coincidentally, neutral-to-alkaline consistent with some amount of burial, and in- sis. Nitrogen-bearing compounds, in both reduced pH conditions also are likely to promote the dependent support for this is allowed (but not and oxidized forms, were measured by SAM formation of clay minerals. Potential authigenic required) by possible incipient chloritization during EGA (52). Evolved NO (mass-to-charge magnetite formation pathways include UV- of mudstone smectite (28). Therefore, it is en- ratio = 30) has an abundance that is substantially promoted reactions with dissolved Fe2+ in the tirely plausible that, collectively, these hundreds above background, pointing to a likely source water column (81) or with carbonate minerals in of meters of strata could represent millions to within the mudstone. Bioavailable N could con- the sediment (42), and partial oxidation of dis- 2+ tens of millions of years. ceivably have been derived from the atmosphere solved Fe by dissolved O2 (80). Perhaps the Episodic drying of this lake environment via fixation by sulfide minerals during reactions most likely scenario is via clay formation and seems certain over this long time scale if not similar to that seen in pyrrhotite, pyrite, and saponitization (28). However the magnetite may over shorter spans represented by thicknesses magnetite assemblages on Earth (69, 70). have formed, the additional presence of Fe3+ equivalent to the exposed Sheepbed mudstone. SAM EGA data from John Klein show car- in nanophase iron oxide indicated by combined However, even if the lake was dry periodically bon to be present, either in crystalline (carbonate?) APXS/CheMin and SAM data (19, 28, 52)pro- or if unconformities interrupted the continuity of phases below 1% abundance, as a carbonate- vides the other half of an iron-based redox cou- fluvial-lacustrine sedimentation, the presence of bearing component in the x-ray amorphous mate- ple in the Sheepbed mudstone. former groundwater—well documented for many rial detected by CheMin, or derived from complex Salinity provides another important con- ancient Mars phenomena (55–57)—may have organic sources not detectable by SAM. Thus, straint on habitability because it affects water provided a refugium to sustain a chemolitho- most or all the major ingredients required to sup- activity, which, when low, can restrict cellular trophic biosphere. Episodic lake rejuvenation port life are present in the Sheepbed mudstone. physiology (82–84). All natural waters contain would permit a subsurface chemoautotrophic As a second step, we identify plausible re- dissolved ions that, through their interactions microbiota to emerge to the surface in response dox couples represented by minerals and other with water molecules, limit the availability of to a rise in groundwater level. compounds preserved in the Sheepbed mud- H2O for hydration reactions. Water activity is We can more specifically assess the habit- stone that might have comprised basic energy- defined by microbiologists as aw = n1/(n1 + n2), ability of the former environment represented by yielding reactions for an elementary microbial where n1 equals moles of water and n2 equals the Sheepbed mudstone from a perspective shaped community (66). The presence of both sulfur and moles of solute (82). Most terrestrial organisms by our knowledge of terrestrial biology. This ap- iron is indicated by ChemCam, APXS, CheMin, cannot survive below aw = 0.9, a few halophilic

www.sciencemag.org SCIENCE VOL 343 24 JANUARY 2014 1242777-11 Exploring Martian Habitability

bacteria grow at aw = 0.85, and some archaea plest interpretation of the sequence of diagenetic 9. A. Fraeman et al., A hematite-bearing layer in Gale crater: Mapping and implications for past aqueous can grow at aw =0.75(83). In the Burns for- events would involve progressive desiccation of — conditions. Geology 41, 1103–1106 (2013). mation at Meridiani Planum, the abundance of mildly saline, pH-neutral waters a very Earth- doi: 10.1130/G34613.1 calcium and magnesium sulfate minerals de- like scenario (92). Such conditions have been 10. This decision was made by the entire MSL Science tected by the Opportunity rover is so great that it envisaged for the very early history of Mars Team with strong consensus to drive a relatively conservatively reconstructs to an ancient water (91), but it is only recently that they have been short distance (~500 m) to examine rocks with a 7 93 94 higher thermal inertia values that also had an activity of w =0.78to0.86,andpossiblyas considered viable for a younger age ( , , ). apparently strong spatial association with the AF low as aw =0.5(85). This shows that water ac- The surprising result is that the stratigraphy of unit. The prediction based on mapping was that tivity was so low during deposition and diagenesis Yellowknife Bay may not only preserve evi- fluvial and/or lacustrine rocks potentially derived of the Burns formation that even though water dence of a habitable environment, but one that is from the crater rim might be encountered. The strata was present, the environment would probably relatively young by martian standards. In the exposed at the base of Mt. Sharp are still considered to be the primary mission objective, and the team have been uninhabitable to all but the hardiest most conservative scenario allowed by geologic has established a departure date from Yellowknife halophiles (86). In contrast, the Sheepbed mud- mapping, the Yellowknife Bay formation repre- Bay of the first week in July 2013. stone provides evidence for both low salinity sents part of the older fill of Gale crater, and 11. The names used for rock and soil targets studied by sediment-transporting water and early diagenetic therefore is roughly Early Hesperian in age (6), Curiosity and reported in this paper are derived from established rock formations of northern Canada. In turn, water, given its very low sulfur and chlorine perhaps overlapping with or postdating times of these names derive from names of local geographic content (19). It is possible that small amounts of bedded sulfate formation elsewhere on Mars. features where they occur, per convention, such as chloride salt were present (19, 28); however, This would indicate that times of sustained surface mountains, streams, lakes, bays, etc. This choice of these were in low enough abundance (<1 to 2%) water, neutral pH, and authigenic clay formation naming scheme celebrates the city of Yellowknife, ’ Northwest Territories, long known to be a port of so as to not substantially reduce water activity extended later into Mars history. The potential- departure for geologic mapping expeditions into some and create a challenge for prokaryotic microbes. ly young age of clay formation (and habitability) of the oldest rocks of North America. Yellowknife Higher salinities are indicated for the waters that does not invalidate the general trend that most crater was designated by the International Astronomical percolated through late diagenetic fractures; rocks that interacted with water early in Mars’ Union, and Yellowknife Bay demarcates the area into which Curiosity descended to study its geologic history. however, these were probably of modest salinity history produced clays, and those that interacted 12. A base map was created based on 25-cm-per-pixel given their exclusively calcium sulfate compo- later produced sulfates. However, much like HiRISE visible image data, 1-m-per-pixel elevation data, sition (85). On Earth, calcium sulfate precipita- Earth’s time-dependent records of iron forma- and 100-m-per-pixel Thermal Emission Imaging System tion is common in evaporitic environments that tion (95, 96) and carbonates (97, 98), it points data that was subdivided into 140 1.2- by 1.2-km thrive with diverse microorganisms (87). Further- to the need to understand those special condi- (0.025°) quadrangles and mapped in detail by more than 30 team members. The group map was compiled, more, there is evidence that halophiles may resist tions, which allow a distinctive aqueous envi- consolidated, and iteratively refined into a single GIS the destructive effects of radiation (88), thought ronment to persist for broad spans of geologic project. to limit habitability in the martian surface envi- time or to recur when the favorable conditions 13. In addition, minor units of eolian fill and/or bedforms ronment (89). again emerge. Curiosity’s detection of a relatively and ejecta blankets adjacent to larger craters are present but are not discussed further. Finally, estimates of pH are similarly impor- young and markedly Earth-like habitable envi- 14. Bradbury Landing refers to the patch of ground where tant for understanding habitability. Again, in the ronment at Gale crater underscores the bio- Curiosity touched down on the surface of Mars. It Burns formation at Meridiani Planum, the pres- logic potential of relatively young fluvio-lacustrine commemorates the noted science fiction writer ence of jarosite indicates low pH with ancient environments. Raymond Bradbury and was designated by NASA. 40 15. E. M. Stolper et al., The petrochemistry of Jake_M: A waters that contained sulfuric acid ( ). How- martian mugearite. Science 341, 1239463 (2013). ever, such low pH does not discount the pres- doi: 10.1126/science.1239463; pmid: 24072927 ence of life; terrestrial analog environments show References and Notes 16. P.-Y. Meslin et al., Soil diversity and hydration as that microbes can survive under these conditions 1. J. P. Grotzinger et al., Mars Science Laboratory Mission observed by ChemCam at Gale crater, Mars. Science and Science Investigation, Mars Science Laboratory 341, 1238670 (2013). doi: 10.1126/science.1238670; because they have adapted their biochemistries Mission and Science Investigation. Space Sci. Rev. 170, pmid: 24072924 to maintain near neutral conditions internally 5–56 (2012). doi: 10.1007/s11214-012-9892-2 17. R. M. E. Williams et al., Martian fluvial conglomerates (90). Nevertheless, the presence of more neutral 2. R. P. I. Irwin III, A. D. Howard, R. A. Craddock, at Gale crater. Science 340, 1068–1072 (2013). waters allows a broader range of microorga- J. M. Moore, An intense terminal epoch of widespread doi: 10.1126/science.1237317; pmid: 23723230 fluvial activity on early Mars: 2. Increased runoff and “ ” nisms to be viable. The presence of clay minerals 18. The morphology of the bay is the result of paleolake development. J. Geophys. Res. 110, E12S15 postdepositional erosion, perhaps by eolian processes, in the Sheepbed mudstone, the absence of any (2005). doi: 10.1029/2005JE002460 as shown by the exposures of stratified rock that evidence for Al mobility, and the absence of 3. M. C. Malin, K. S. Edgett, Sedimentary rocks of record dissection. These are eroded rocks and not, iron sulfate minerals, even in the sulfate-filled early Mars. Science 290, 1927–1937 (2000). for example, shoreline berms. Therefore, the late diagenetic veins, indicates a relatively neu- doi: 10.1126/science.290.5498.1927; pmid: 11110654 present-day topography does not reflect that 4. R. B. Anderson, J. F. Bell III, Geologic mapping and at the time of deposition of those strata. tral pH. This neutral pH could have existed for characterization of Gale crater and implications for 19. S. M. McLennan et al., Elemental geochemistry of all phases of aqueous history, from the primary its potential as a Mars Science Laboratory landing site. sedimentary rocks at Yellowknife Bay, Gale crater, Mars. waters of deposition that may have created a The Mars Journal 5,76–128 (2010). doi: 10.1555/ Science 343, 1244734 (2014). body of standing water, to early diagenesis, and mars.2010.0004 20. S. R. Taylor, S. M. McLennan, Planetary Crusts: 5. R. E. Milliken, D. L. Bish, Sources and sinks of clay Their Composition, Origin, and Evolution (Cambridge through the close of late diagenesis. minerals on Mars. Philos. Mag. 90, 2293–2308 (2010). Univ. Press, Cambridge, 2009). At Yellowknife Bay, there is no hint of the doi: 10.1080/14786430903575132 21. M. E. Schmidt et al., APXS of first rocks encountered strongly acidic conditions that have been thought 6. B. J. Thomson et al., Constraints on the origin by Curiosity in Gale crater: Geochemical diversity to especially describe the planet’s younger his- and evolution of the layered mound in Gale and volatile element (K and Zn) enrichment. crater, Mars using Mars Reconnaissance Orbiter Lunar Planet. Sci. Conf. 44, abstract 1278 (2013). tory of aqueous alteration, sedimentation, and – 74 83 91 data. Icarus 214, 413 432 (2011). doi: 10.1016/ 22. N. D. Smith, Some sedimentological aspects of habitability ( , , ). The record of aqueous j.icarus.2011.05.002 planar cross-stratification in a sandy braided river. activity at Yellowknife Bay is likely prolonged 7. J. P. Grotzinger, R. E. Milliken, The sedimentary rock J. Sediment. Petrol. 42, 624–634 (1972). and complex, involving several stages of di- record of Mars: Distribution, origins, and global 23. G. A. Macdonald, Pahoehoe, aa, and block lava. agenesis, including clay formation, that culmi- stratigraphy, in Sedimentary Geology of Mars, Am.J.Sci.251, 169–191 (1953). doi: 10.2475/ J. Grotzinger, R. E. Milliken, Eds. (SEPM Special nate with precipitation of calcium sulfate salts ajs.251.3.169 Publication, Tulsa, OK, 2012), vol. 102. 24. A. Hurst, A. Scott, M. Vigorito, Physical characteristics of in veins, but without attendant indicators of 8. J. P. Grotzinger, Beyond water on Mars. Nat. Geosci. 2, sand injectites. Earth Sci. Rev. 106, 215–246 (2011). acidic waters such as iron sulfates. The sim- 231–233 (2009). doi: 10.1038/ngeo480 doi: 10.1016/j.earscirev.2011.02.004

1242777-12 24 JANUARY 2014 VOL 343 SCIENCE www.sciencemag.org RESEARCH ARTICLE

25. J. A. Ross, J. Peakall, G. M. Keevil, An integrated model of 47. M. D. Pollock, L. C. Kah, J. K. Bartley, Morphology of 68. M. S. Milman-Barris et al., Zoning of phosphorous in extrusive sand injectites in cohesionless sediments. molar-tooth structures in Precambrian carbonates: igneous olivine. Contrib. Mineral. Petrol. 155, 739–765 Sedimentology 58, 1693–1715 (2011). doi: 10.1111/ Influence of substrate rheology and implications for (2008). doi: 10.1007/s00410-007-0268-7 j.1365-3091.2011.01230.x genesis. J. Sediment. Res. 76, 310–323 (2006). 69. M. A. Schoonen, Y. Xu, Nitrogen reduction under 26. M. El-Tabakh, R. Riccioni, B. C. Schreiber, Evolution of doi: 10.2110/jsr.2006.021 hydrothermal vent conditions: Implications for the late Triassic rift basin evaporites (Passaic Formation): 48. A. Hurst, J. A. Cartwright, D. Duranti, in Subsurface prebiotic synthesis of C-H-O-N compounds. Astrobiology Newark Basin, Eastern North America. Sedimentology Sediment Mobilisation, R. R. H. P. Van Rensbergen, 1, 133–142 (2001). 44, 767–790 (1997). doi: 10.1046/j.1365- A. J. Maltman, C. K. Morley, Ed. (Geological Society of London 70. D. P. Summers, R. C. B. Basa, B. Khare, D. Rodoni, 3091.1997.d01-47.x Special Publication, London, 2003), vol. 216, pp. 123–137. Abiotic nitrogen fixation on terrestrial planets: 27. J. A. Fisher, G. J. Nichols, D. A. Waltham, Unconfined 49. I. A. Kane, Development and flow structures of sand Reduction of NO to ammonia by FeS. Astrobiology 12, flow deposits in distal sectors of fluvial distributary injectities: The Hind Sandstone Member injectite complex, 107–114 (2012). doi: 10.1089/ast.2011.0646; systems: Examples from the Miocene Luna and Huesca Carboniferous, UK. Mar. Pet. Geol. 27, 1200–1215 pmid: 22283408 Systems, northern Spain. Sediment. Geol. 195,55–73 (2010). doi: 10.1016/j.marpetgeo.2010.02.009 71. V. I. Oyama, B. J. Berdahl, G. C. Carle, Preliminary (2007). doi: 10.1016/j.sedgeo.2006.07.005 50. M. Huuse, J. Cartwright, A. Hurst, N. Steinsland, in Sand findings of the Viking gas exchange experiment and a 28. D. T. Vaniman et al., Mineralogy of a mudstone at Injectities: Implications for Hydrocarbon Exploration and model for martian surface chemistry. Nature 265, Yellowknife Bay, Gale crater, Mars. Science 343, Production,A.Hurst,J.Cartwright,Eds.(American 110–114 (1977). doi: 10.1038/265110a0 1243480 (2014). Association of Petroleum Geologists Memoir, Tulsa, OK, 72. H. P. Klein, The Viking Biological Investigation, General 29. P. E. Potter, J. B. Maynard, P. Depetris, Mud and 2007), vol. 87. aspects. J. Geophys. Res. 82, 4677–4680 (1977). Mudstones (Springer, Berlin, 2005). 51. J. W. Cosgrove, Hydraulic fracturing during the formation doi: 10.1029/JS082i028p04677 30. P. B. Attewell, I. W. Farmer, Principles of Engineering and deformation of a basin: A factor in the dewatering of 73. M. H. Hecht et al., Detection of perchlorate and the Geology (Chapman and Hall, London, New York, 1976). low-permeability sediments. AAPG Bull. 85, 737–748 soluble chemistry of martian soil at the Phoenix lander 31. Supplementary materials are available on Science Online. (2001). site. Science 325,64–67 (2009). pmid: 19574385 32. D. Winston, Fluvial Systems of the Precambrian Belt 52. D. W. Ming et al., Volatile and organic compositions of 74. S. W. Squyres, A. H. Knoll, Sedimentary rocks and Supergroup, Montana and Idaho, U.S.A. Fluvial sedimentary rocks in Yellowknife Bay, Gale crater, Mars. Meridiani Planum: Origin, diagenesis, and implications Sedimentology Memoir 5, 343 (1977). Science 343, 1245267 (2014). for life of Mars. Earth Planet. Sci. Lett. 240,1–10 33. D. S. McCormick, J. P. Grotzinger, Distinction of 53. J. D. Collinson, Sedimentology of unconformities within (2005). doi: 10.1016/j.epsl.2005.09.038 marine from alluvial Facies in the Paleoproterozoic a fluvio-lacustrine sequence; Middle Proterozoic of 75. L. A. Leshin et al., Volatile, isotope, and organic (1.9 Ga) burnside formation, Kilohigok basin, Eastern North Greenland. Sediment. Geol. 34, 145–166 analysis of martian fines with the Mars Curiosity N.W.T., Canada. J. Sediment. Petrol. 63, 398–419 (1983). doi: 10.1016/0037-0738(83)90084-2 rover. Science 341, 1238937 (2013). (1993). 54. L. B. Aspler, J. R. Chiarenzelli, B. L. Cousens, Fluvial, doi: 10.1126/science.1238937; pmid: 24072926 34. A. S. Yen et al., An integrated view of the chemistry and lacustrine and volcanic sedimentation in the Angikuni 76. R. E. Milliken et al., Opaline silica in young deposits mineralogy of martian soils. Nature 436,49–54 (2005). sub-basin, and initiation of ∼1.84–1.79 Ga Baker on Mars. Geology 36, 847–850 (2008). doi: 10.1130/ doi: 10.1038/nature03637; pmid: 16001059 Lake Basin, western Churchill Province, Nunavut, G24967A.1 35. P. R. Christensen, Regional dust deposits on Mars: Canada. Precambrian Res. 129, 225–250 (2004). 77. R. V. Morris et al., Mössbauer mineralogy of rock, Physical properties, age, and history. J. Geophys. Res. doi: 10.1016/j.precamres.2003.10.004 soil, and dust at Gusev crater, Mars: Spirit’s journey Solid Earth 91, 3533–3545 (1986). doi: 10.1029/ 55. H. H. Kieffer, B. M. Jakosky, C. Snyder, M. Matthews, through weakly altered olivine basalt on the plains JB091iB03p03533 Mars,H.H.Kieffer,B.M.Jakosky,C.Snyder,M.Matthews, and pervasively altered basalt in the Columbia Hills. 36. N. T. Bridges, D. R. Muhs, in Sedimentary Geology Eds. (Univ. of Arizona Press, Tucson, AZ, 1992). J. Geophys. Res. Planets 111, E02S13 (2006). of Mars (SEPM Special Publication, Tulsa, OK, 2012), 56. J. F. Bell III, The Martian Surface: Composition, doi: 10.1029/2005JE002584 vol. 102, pp. 169–182. Mineralogy, and Physical Properties, J. F. Bell III, Ed. 78. R. V. Morris et al., Mossbauer mineralogy of rock, soil, and 37. N. Mangold, V. Ansan, P. Masson, C. Vincendon, (Cambridge Univ. Press, New York, 2008). dust at Meridiani Planum, Mars: Opportunity's journey Estimate of aeolian dust thickness in Arabia Terra, Mars: 57. J. C. Andrews-Hanna, K. W. Lewis, Early Mars hydrology: across sulfate-rich outcrop, basaltic sand and dust, and Implications of a thick mantle (>20 m) for hydrogen 2. Hydrological evolution in the Noachian and hematite lag deposits. J. Geophys. Res. 111, E12S15 (2006). detection. Geomorphol.: Relief, Processes, Environ. Hesperian epochs. J. Geophys. Res. Planets 116, E02007 79. J. F. Bell 3rd et al., Pancam multispectral imaging 1/2009,23–32 (2009). (2011). doi: 10.1029/2010JE003709 results from the Opportunity Rover at Meridiani Planum. 38. P. M. Sadler, Sediment accumulation rates and the 58. C. R. Woese, G. E. Fox, Phylogenetic structure of the Science 306, 1703–1709 (2004). doi: 10.1126/ completeness of stratigraphic sections. J. Geol. 89, prokaryotic domain: The primary kingdoms. Proc. Natl. science.1105245; pmid: 15576603 569–584 (1981). doi: 10.1086/628623 Acad. Sci. U.S.A. 74, 5088–5090 (1977). doi: 10.1073/ 80. R. M. Cornell, U. Schwertmann, The Iron Oxides: 39. A. M. Sarna-Wojcicki, S. Shipley, R. B. Waitt, D. Dzurisin, pnas.74.11.5088; pmid: 270744 Structure, Properties, Reactions, Occurrence and Uses S. H. Wood, in The 1980 Eruptions of Mount St. Helens, 59. E. I. Friedmann, Endolithic microorganisms in the (Wiley-VCH, Weinheim, Germany, ed. 2, 2003). Washington, P. W. Lipman, D. R. Mullineaux, Eds. antarctic cold desert. Science 215, 1045–1053 81. G. N. Schrauzer, T. D. Guth, Hydrogen evolving systems. (U.S. Geological Survey Professional Paper, Reston, VA, (1982). doi: 10.1126/science.215.4536.1045; 1. The formation of molecular hydrogen from aqueous 1981), vol. 1250, pp. 577–600. pmid: 17771821 suspensions of iron(II) hydroxide and reactions with 40. S. M. McLennan et al., Provenance and diagenesis of the 60. K. O. Stetter, Extremophiles and their adaptation to reducible substrates, including molecular nitrogen. evaporite-bearing Burns formation, Meridiani Planum, hot environments. FEBS Lett. 452,22–25 J. Am. Chem. Soc. 98,3508–3513 (1976). Mars. Earth Planet. Sci. Lett. 240,95–121 (2005). (1999). doi: 10.1016/S0014-5793(99)00663-8; doi: 10.1021/ja00428a019 doi: 10.1016/j.epsl.2005.09.041 pmid: 10376671 82. W. D. Grant, Life at low water activity. Philos. Trans. 41. P. Fralick, J. P. Grotzinger, L. Edgar, in Sedimentary 61. M. Rossi et al., Extremophiles 2002. J. Bacteriol. R. Soc. London Ser. B 359, 1249–1267 (2004). Geology of Mars, J. P. Grotzinger, R. E. Milliken, Eds. 185, 3683–3689 (2003). doi: 10.1128/JB.185.13. doi: 10.1098/rstb.2004.1502; pmid: 15306380 (SEPM Special Publication, Tulsa, OK, 2012), vol. 102. 3683-3689.2003; pmid: 12813059 83. A. H. Knoll, J. Grotzinger, Water on Mars and the 42. J. D. Kim, N. Yee, V. Nanda, P. G. Falkowski, Anoxic 62. K. H. Nealson, Sediment bacteria: Who’s there, what prospect of martian life. Elements 2, 169–173 (2006). photochemical oxidation of siderite generates molecular are they doing, and what’s new? Annu. Rev. Earth doi: 10.2113/gselements.2.3.169 hydrogen and iron oxides. Proc. Natl. Acad. Sci. U.S.A. Planet. Sci. 25, 403–434 (1997). doi: 10.1146/ 84. J. P. Williams, J. E. Hallsworth, Limits of life in hostile 110, 10073–10077 (2013). doi: 10.1073/ annurev.earth.25.1.403; pmid: 11540735 environments: No barriers to biosphere function? pnas.1308958110; pmid: 23733945 63. H. G. Schlegel, General Microbiology (Cambridge Univ. Environ. Microbiol. 11, 3292–3308 (2009). doi: 10.1111/ 43. G. Furniss, J. F. Rittel, D. Winston, Gas bubble and Press, Cambridge, ed. 7, 1993). j.1462-2920.2009.02079.x; pmid: 19840102 expansion crack origin of “molar-tooth” calcite structures 64. G. Gottschalk, Microbial Metabolism (Springer, New York, 85. N. J. Tosca, A. H. Knoll, S. M. McLennan, Water activity in the Middle Porterozoic Belt Supergroup, western 1994). and the challenge for life on early Mars. Science 320, Montana. J. Sediment. Res. 68, 104–114 (1998). 65. R. J. Parkes et al., Deep sub-seafloor prokaryotes 1204–1207 (2008). doi: 10.1126/science.1155432; doi: 10.2110/jsr.68.104 stimulated at interfaces over geological time. Nature pmid: 18511686 44. The term “raised ridges” may seem redundant, but the 436, 390–394 (2005). doi: 10.1038/nature03796; 86. S. Fendrihan et al., Spherical particles of halophilic designation “raised” is meant to denote the substantial pmid: 16034418 archaea correlate with exposure to low water height of the ridges compared to their width. 66. K. H. Nealson, P. G. Conrad, Life: past, present and activity—implications for microbial survival in fluid 45. P. S. Plummer, V. A. Gostin, Shrinkage cracks: future. Philos. Trans. R. Soc. London Ser. B 354, inclusions of ancient halite. Geobiology 10, 424–433 Desiccation or synaeresis? J. Sediment. Res. 51, 1923–1939 (1999). pmid: 10670014 (2012). doi: 10.1111/j.1472-4669.2012.00337.x; 1147–1156 (1981). 67. D. Z. Sousa et al., Activity and viability of methanogens pmid: 22804926 46. B. R. Pratt, Syneresis cracks: Subaqueous shrinkage in in anaerobic digestion of unsaturated and saturated 87. C. Demergasso et al., Distribution of prokaryotic argillaceous sediments caused by earthquake-induced long-chain fatty acids. Appl. Environ. Microbiol. 79, genetic diversity in athalassohaline lakes of the Atacama dewatering. Sediment. Geol. 117,1–10 (1998). 4239–4245 (2013). doi: 10.1128/AEM.00035-13; Desert, Northern Chile. FEMS Microbiol. Ecol. 48,57–69 doi: 10.1016/S0037-0738(98)00023-2 pmid: 23645196 (2004). doi: 10.1016/j.femsec.2003.12.013;pmid:19712431

www.sciencemag.org SCIENCE VOL 343 24 JANUARY 2014 1242777-13 Exploring Martian Habitability

88. A. Kish et al., Salt shield: Intracellular salts provide Mars Reconnaissance Orbiter. J. Geophys. Res. Planets Acknowledgments: We are indebted to the MSL Project cellular protection against ionizing radiation in 114, E00D06 (2009). doi: 10.1029/2009JE003342 engineering and management teams for their exceptionally the halophilic archaeon, Halobacterium salinarum 94. A. G. Fairén et al., Noachian and more recent skilled and diligent efforts in making the mission as effective NRC-1. Environ. Microbiol. 11, 1066–1078 (2009). phyllosilicates in impact craters on Mars. Proc. Natl. as possible and enhancing science operations. We are also doi: 10.1111/j.1462-2920.2008.01828.x; Acad. Sci. U.S.A. 107, 12095–12100 (2010). grateful to all those MSL Science Team members who pmid: 19452594 doi: 10.1073/pnas.1002889107; pmid: 20616087 participated in tactical and strategic operations. Without the 89. L. R. Dartnell, L. Desorgher, J. M. Ward, A. J. Coates, 95. H. L. James, A. F. Trendall, Banded iron formation: support of both the engineering and science teams, the data Modelling the surface and subsurface martian Distribution in time and paleoenvironmental significance, presented here could not have been collected. Some of this radiation environment: Implications for astrobiology. Mineral Deposits and the Evolution of the Biosphere: research was carried out at the Jet Propulsion Laboratory, Geophys. Res. Lett. 34, L02207 (2007). Dahlem Workshop Report, H. D. Holland, A. M. Schidlowski, California Institute of Technology, under a contract with NASA. doi: 10.1029/2006GL027494 Eds. (Springer, Berlin, Heidelberg, New York, 1982), Data presented in this paper are archived in the Planetary Data 90. A. I. López-Archilla, I. Marin, R. Amils, Microbial vol. 3. System (pds.nasa.gov). community composition and ecology of an acidic aquatic 96. A. Bekker et al., Iron formation: The sedimentary environment: The Tinto River, Spain. Microb. Ecol. 41, product of a complex interplay among mantle, 20–35 (2001). pmid: 11252161 tectonic, oceanic, and biospheric processes. Econ. Supplementary Materials 91. J.-P. Bibring et al., Global mineralogical and Geol. 105, 467–508 (2010). doi: 10.2113/ www.sciencemag.org/content/343/6169/1242777/suppl/DC1 aqueous mars history derived from OMEGA/Mars gsecongeo.105.3.467 Supplementary Text Express data. Science 312, 400–404 (2006). 97. J. P. Grotzinger, in Controls on Carbonate Platforms and Figs. S1 to S4 doi: 10.1126/science.1122659; pmid: 16627738 Basin Development, P. D. Crevello, J. L. Wilson, J. F. Sarg, Tables S1 and S2 92. T. F. Bristow, R. E. Milliken, Terrestrial perspective on J. F. Read, Eds. (SEPM Special Publication, Tulsa, OK, References (99, 100) authigenic clay mineral production in ancient 1989), vol. 44, pp. 79–106. MSL Science Team Author List martian lakes. Clays Clay Miner. 59, 339–358 (2011). 98. J. P. Grotzinger, N. P. James, in Carbonate Sedimentation doi: 10.1346/CCMN.2011.0590401 and Diagenesis in the Evolving Precambrian World, 4 July 2013; accepted 6 November 2013 93. S. L. Murchie et al., A synthesis of martian aqueous J. P. Grotzinger, N. P. James, Eds. (SEPM Special Published online 9 December 2013; mineralogy after 1 Mars year of observations from the Publication, Tulsa, OK, 2000), vol. 67, pp. 3–20. 10.1126/science.1242777

1242777-14 24 JANUARY 2014 VOL 343 SCIENCE www.sciencemag.org